1. Field of the Invention
The present invention relates to a semiconductor photoelectric conversion device which has a laminate member composed of a plurality of PIN structures formed one on the other. Further, the invention pertains to a method for the manufacture of such a semiconductor photoelectric conversion device.
2. Description of the Prior Art
Heretofore a tandem type semiconductor photoelectric conversion device has been proposed which has a laminate member comprised of at least two first and second PIN structures, with the first PIN structure disposed on the light impinging side of the device.
In this case, the first and second PIN structures each have such a construction that a first conductivity type (P- or N-type) non-single-crystal semiconductor layer, an I-type non-single-crystal semiconductor layer and a second conductivity type (reverse from the first one) non-single-crystal semiconductor layer are laminated in that order or in the reverse order. The I-type layers of the first and second PIN structures are doped with a recombination center neutralizer such as hydrogen or a halogen. The I-type layer of the first PIN structure disposed on the light impinging side of the device has a larger optical energy gap than does the I-type layer of the second PIN structure.
With such a tandem type semiconductor photoelectric conversion device, the I-type layers of the first and second PIN structures are excited by the incidence of light to the laminate member from the side of the first PIN structure, by which carriers, i.e. electron-hole pairs, are generated in each I-type layer, and the electrons and holes respectively flow into one of the first and second conductivity type layers, that is, the N-type layer and into the other, that is, the P-type layer, developing photo voltage.
Since the tandem type semiconductor photoelectric conversion device has the construction that the first and second PIN structures are electrically connected in series, it is possible to convert light to electric power which has a voltage about twice as high as that obtainable with a non-tandem type semiconductor photoelectric conversion device which has one PIN structure similar to those of the above tandem type device.
In the tandem type photoelectric conversion device, the I-type layers of the first and second PIN structures are excited by the incidence of light. In this case, the I-type layers are each most sensitive to light of a wavelength-corresponding to its energy gap. Letting the energy gaps of the I-type layers of the first and second PIN structures be represented by Eg.sub.1 and Eg.sub.2 (where Eg.sub.1 &gt;Eg.sub.2), respectively, and the wavelengths of light corresponding to the energy gaps Eg.sub.1 and Eg.sub.2 by .lambda..sub.1 and .lambda..sub.2 (where .lambda..sub.1 &lt;.lambda..sub.2), respectively, the I-type layer of the first PIN structure is excited most by light of the wavelength .lambda..sub.1 and the I-type layer of the second PIN structure is excited most by light of the wavelength .lambda..sub.2.
Accordingly, the tandem type semiconductor photoelectric conversion device has another advantage that it is able to convert light to electric power over a wide wavelength range as compared with the on-tandem type device.
In the conventional tandem type semiconductor photoelectric conversion device, however, the first and second PIN structures are formed by the CVD method in that order or in the reverse order. Accordingly, the second (or first) conductivity type layer of the first PIN structure and the first (or second) conductivity type layer of the second PIN structure are formed by the CVD method in that order or in the reverse order.
In consequence, regions in which first and second conductivity type impurities are present as a mixture are formed respectively in second (or first) conductivity type layer of the first PIN structure on the side of its second PIN structure and in first (or second) conductivity type layer of the second PIN structure on the side of the its first PIN structure. In such regions the first and second conductivity type impurities are compensated for each other. These regions are low in conductivity and high in light absorptivity than the second (or first) conductivity type layer of the first PIN structure and the first (or second) conductivity type layer of the second PIN structure. Therefore, the first and second PIN structures are not connected in tandem without making electrically good ohmic contact with each other. This will cause internal electrical losses and inflict appreciable losses on light incident to the I-type layer of the second PIN structure through the first PIN structure, resulting in the reduction of the photoelectric conversion efficiency.
In the conventional tandem type semiconductor photoelectric conversion device, the I-type layers of the first and second PIN structures are both formed of an amorphous semiconductor. On account of this, the mobility of carriers which are generated in the I-type layers by incidence of light thereto is lower than in the case where the I-type layers are crystallized. Further, the degree of recombination of the carriers in the I-type layers of the first and second PIN structures is higher than in the case where the I-type layers are crystallized.
Accordingly, the prior art tandem type semiconductor photoelectric conversion device possesses the defect that its photoelectric conversion rate and efficiency are both relatively low.
In the conventional tandem type semiconductor photoelectric conversion device, the first and second PIN structures have different semiconductor compositions so that the I-type layer of the first PIN structure has a larger optical energy gap than does the I-type layer of the second PIN structure. That is, for example, the I-type layer of the first PIN structure is formed of amorphous silicon, whereas the I-type layer of the second PIN structure is formed of amorphous Si.sub.x Ge.sub.1-x (where 0&lt;x&lt;1).
This means that an inexpensive material cannot be used for the I-type layers of the first and second PIN structures. Hence, the conventional semiconductor photoelectric conversion device is high in manufacturing costs. This is even more marked in the case of using amorphous silicon for the I-type layer of the first PIN structure and amorphous Si.sub.x Ge.sub.1-x for the I-type layer of the second PIN structure, as mentioned above. The reason for this is as follows: The I-type layer of the first PIN structure can be formed by a CVD method using silane (SiH.sub.4), and the I-type layer of the second PIN structure can be formed by a CVD method using germane GeH.sub.4). Although silicon forming the silane (SiH.sub.4) is available at low cost, germanium forming the germane (GeH.sub.4) is costly, resulting in an increase in the manufacturing costs of the semiconductor photoelectric conversion device.